Calculating momentum from maximum threat

53:47threatening situations in that minute. They're not actually doing a average over all the passes. They're looking at the extrema. Like, what was the most dangerous thing that team A did versus what was the most dangerous thing that B did? >> minute interval. >> Yeah. That's kind of nice because now you're not washing out the stuff that actually matters. Because in football, like things can happen very quickly, right? So, you want to reward the maximum stuff more than like oh, just passes here and there. If you just average, you're going to like water down everything. So, that's one thing that I thought was really cool. And a lot of times in data analytics, you have to make these kinds of choices about what is the part of the data that you care about. So, they measure the possession

54:27value of that maximum move over the past minute, and they weight it based on how long ago that happened. So, they've got like some time kernel that they convolve with. >> That's actually a good way to do it. >> Because if you just did it on like raw possession stats, like you know, you could have even in the Belgium-Iran game, right? Belgium would be way over weighted, and they were in large pockets where they were not particularly dangerous. And Iran had these like, you know, moments that were much more um impactful even though they were mostly out of possession at least in the first half. >> Yeah, exactly. That's why I think I like that they they only keep these like the the maximum part instead of averaging. And then the the momentum is given by

55:08the difference between the two. >> Yeah. Okay. >> And one really cool sort of visualization that I saw was I mean, there was a tweet by someone saying that like the match hydration breaks